CLIO QC EXPLAINED WITH APPS - Audiomatica · CLIO QC EXPLAINED WITH APPS ... The QC software is a...

AN-009APPLICATION NOTE

CLIO QC EXPLAINED WITH APPSby Audiomatica – [email protected]

CONTENTS

1 FEATURES OF CLIO QC............................................................................. 1.1 THE OPERATOR'S POINT OF VIEW........................................................ 1.2 THE ENGINEER'S POINT OF VIEW........................................................ 1.3 THE COMPANY'S POINT OF VIEW.........................................................

2 THE QC SOFTWARE OPERATION................................................................. 2.2 THE CYCLIC QC SCRIPT...................................................................... 2.3 THE REFERENCE FILE......................................................................... 2.4 THE LIMITS FILE................................................................................ 2.5 LIMITS FILE FOR TWO CHANNELS STEREO OPERATION...........................

3 THE CLIO QC TCP/IP MEASUREMENT SERVER.............................................. 3.1 INVOKING THE QC SERVER................................................................. 3.2 CONNECTING TO THE QC SERVER........................................................ 3.3 INTERACTING WITH THE QC SERVER................................................... 3.4 NOTES ABOUT QC SERVICES...............................................................

4 HANDS ON QC......................................................................................... 4.1 WHAT TO KNOW ABOUT QC SCRIPTS................................................... 4.2 HOW TO WRITE MY FIRST QC SCRIPT...................................................

5 RUNNING A QC SCRIPT............................................................................. 5.1 DESKTOP AND WINDOWS MANAGEMENT.............................................. 5.2 THE QC RESULT PANEL....................................................................... 5.3 THE QC BANNER................................................................................ 5.4 THE QC REPORT PANEL....................................................................... 5.5 REVIEWING A MEASUREMENT............................................................. 5.6 THE SKIP LAST BUTTON......................................................................

6 NOTES ON LIMITS CURVES......................................................................... 6.1 ABSOLUTE VS. RELATIVE FREQUENCY LIMITS........................................ 6.2 AVERAGE LEVEL CHECK...................................................................... 6.3 ALIGNED MASK................................................................................. 6.4 SENSITIVITY CHECK.......................................................................... 6.5 FLOATING LIMITS VS. FLOATING CURVES............................................. 6.6 SINUSOIDAL A/B STEREO DIFFERENCE CHECK...................................... 6.7 SINUSOIDAL THD AND FAST-TRACK RUB&BUZZ CHECK.......................... 6.8 THIELE&SMALL PARAMETERS CHECK.................................................... 6.9 LOUDNESS RATING CALCULATION AND CHECK...................................... 6.10 MULTIMETER LIMITS FILES................................................................

7 MANAGING PRODUCTION BATCHES............................................................ 7.1 DIRECTORIES CREATED BY CLIO QC.................................................... 7.2 PRODUCTION REPORT FILES............................................................... 7.3 AUTOSAVED DATA FILES..................................................................... 7.4 STATISTICAL INFORMATION ON MEASURED DATA.................................. 7.5 SERIAL NUMBER MANAGEMENT...........................................................

8 INTERACTING WITH EXTERNAL HARDWARE................................................. 8.1 INPUT SENSITIVITY AND OUTPUT VOLTAGE CONTROL............................ 8.2 QCBOX MODEL 5 DC OUTPUT CONTROL................................................

Rev. 04/14 www.audiomatica.com

AUDIOMATICA

mailto:[email protected]

CLIO QC EXPLAINED WITH APPS

8.3 CLIOQC AMPLIFIER&SWITCHBOX CONTROL.......................................... 8.4 EXTERNAL TRIGGER........................................................................... 8.5 TTL SIGNALS GENERATION................................................................. 8.6 TIME DELAYS GENERATION................................................................. 8.7 PARALLEL PORT SIGNALS MANAGEMENT............................................... 8.8 QCBOX MODEL 5 DIGITAL I/O SIGNALS MANAGEMENT.......................... 8.9 RS-232 SERIAL PORT CONTROL...........................................................

10 COMPLETE EXAMPLE: FAST, SINGLE-TEST LOUDSPEAKER QC....................... 10.1 HARDWARE REQUIRED..................................................................... 10.2 MEASURING THE REFERENCE FREQUENCY RESPONSE........................... 10.3 MEASURING THE REFERENCE IMPEDANCE RESPONSE........................... 10.4 INTEGRATING THE QC REFERENCE FILE.............................................. 10.5 PROGRAMMING THE QC SCRIPT......................................................... 10.6 RUNNING THE QC TEST.................................................................... 10.7 ADDING THE INTERFACE TO AUTOMATION..........................................

11 QC APPS............................................................................................... 11.1 QC OF A MICROPHONE PREAMPLIFIER................................................ 11.2 THE AMPLIFIER&SWITCHBOX UNDER QC............................................. 11.3 A TEST ON A STEREO ELECTRONIC EQUIPMENT................................... 11.4 A CYCLIC SCRIPT (USED TO MANAGE MY ROGERS LS3/5A TWO-WAY LOUDSPEAKER PRODUCTION).................................................................... 11.5 QC OF A TELEPHONE WITH LOUDNESS RATING CHECK......................... 11.6 ON RUB & BUZZ DETECTION (1)........................................................ 11.7 ON RUB & BUZZ DETECTION (2)........................................................ 11.8 A C++ CLIENT APPLICATION TO CONNECT TO TCP/IP SERVER...............

2/85 www.audiomatica.com


1 FEATURES OF CLIO QC

CLIO QC is exceptionally powerful as it relies on the power of CLIO. Here is a list ofthe parameters that can be calculated within each measurement:

Sinusoidal - Frequency response and impedance response (mono or stereo tests)- Average (or single frequency) level- Sensitivity (average or up to eight frequencies)- Polarity- Total harmonic distortion response- Single harmonic response (from 2nd to 10th)- Fast-Track Rub&Buzz response- T&S parameters (Fs,Qt,Qe,Qm,Cms,Mms,Mmd,Vas,Bl,dBSPL,ZMin)- Loudness Rating (RLR, SLR, STMR)

MLS&CHIRP- Frequency response or impedance response (mono tests) - Average (or single frequency) level- Sensitivity (average or up to eight frequencies) - Polarity- T&S parameters (Fs,Qt,Qe,Qm,Cms,Mms,Mmd,Vas,Bl,dBSPL,ZMin)- Loudness Rating (RLR, SLR, STMR)

FFT - Frequency response with definable stimulus (mono tests, alsointeractive)- Average (or single frequency) level- Sensitivity (average or up to eight frequencies)

METER - SPL, Volts, THD, IMD single parameter (mono tests, also interactive)

The QC processor is able of handling a virtually unlimited sequence of tests toaccomplish even the most complex tasks; on the other hand a single ultra-fastsinusoidal test may ensure you production cycle times of less than 1 secondwith total integration with the line controller.

Some of the QC management features are better explained starting from thevarious people taking part in this complex operation and their points of view:

- The operator working on the line- The quality control engineer responsible for production line operation- The company and its managers controlling the overall process

All QC operations can be password protected; file operation can be restricted bytheir digital signature.



1.1 THE OPERATOR'S POINT OF VIEW

A quality control test can be controlled by simple Go-NoGo masks letting even theleast experienced operator work without problems and with no learning curve.

Figure 1

A more complex operation foresees the continuous display of the measurementsexecuted until the reaching of the final result.

Figure 2

A third possibility is to view and interact with the test sequence during its



execution.

Figure 3

Completed test information and reports are always presented to the user.

Figure 4



1.2 THE ENGINEER'S POINT OF VIEW

As the QC is integrated inside the CLIO software no new user interface has to belearned by the engineer who has experience of CLIO inside her or his researchlaboratory. A quality control test relies on real measurements saved on disk and ona simple text script.

Figure 5

Defining a QC script is easy as it requires the writing only a few descriptive lines oftext, no programming languages or complex instructions are involved.

Figure 6

It is possible to capture the active measurement; the check masks can also beinput in a visual manner drawing limits over the measurement; debugging is helpedby an internal corrector.

Figure 7



1.3 THE COMPANY'S POINT OF VIEW

CLIO when used for quality control executes line testing in a fast, accurate andreliable manner. Its flexibility permits easy handling of trade-offs betweenparameters like speed and accuracy always matching the company's’ needs. Theautosaving and exporting capabilities together the complete result reporting givesinstant access to the production parameters and statistics even during its operation.The production batch is fully managed while preserving serial number coherence.

Figure 8

Figure 9



2 THE QC SOFTWARE OPERATION

The QC software is a "file driven" event processor that, in sequence, performs anumber of user-defined measurements to test the quality of a production line.The text file ('.qc' extension) driving this process is called the QC Script.

The QC script stores information in logical groupings, called sections, initiated by abracketed keyword in the form [keyword].As an example this is a script composed by two sections, one defining globalvariables, the second defining an MLS & LogChirp measurement:

[GLOBALS]COMPANY=MY COMPANYTITLE=MY QUALITY CONTROLBATCH=MY PRODUCTION BATCH NAME

[MLS]OUT=1.000 VINA=0INB=0REFERENCE=MYREFERENCE.MLSLIMITS=MYLIMIT.LIM

CLIO's QC processor does the following job:

- reads the QC script and loads it in memory- interprets it- executes all the tests - reports the test result and production statistics- manages the production batch and serial number- prompts for the next test

The following block diagram outlines the QC process.

WAIT FOR USEROR TRIGGER TOSTART QC TEST

REPORT,STATISTICS

& BATCHMANAGEMENT

MEASUREMENTSENDED ?

ALL RESULTSGOOD ?

QC TEST

GOOD

QC TEST

BAD

LOAD QCSCRIPT

YES

YES

NONO

PERFORMMEASUREMENT

Figure 10

You can see the operation of loading the QC script from disk that begins our qualitycontrol session; then CLIO waits for that the user, or an external trigger (forexample a TTL signal from the automation controller), to give the actual start to theQC test; the measurements defined are then executed in sequence until the last isreached; the result of the test is given by the sum of all the checks done inside thetest sequence, it is only good if all checks gave a positive result; the QC test ends



by updating the report and statistics while managing the production batch; the nextdevice can then be put under test.

To proceed further it is advisable to go into the former block diagram in greaterdetail; this is done in Fig.11 and 12; Fig.11 zooms the entire QC test sequenceadding the blocks in red, while Fig.12 zooms the "Perform Measurement " singleblock (the blue one).

WAIT FOR USER OR TRIGGER TO START QC TEST

PERFORM MEASUREMENT

MEASUREMENTS ENDED ?

ALL RESULTS GOOD ?

QC TEST

GOOD

QC TEST

BAD

LOAD QC SCRIPT

YES

YES

NO

NO

SHOW MEASUREMENT

IF DISPLAYONBAD

SHOW AND PROMPT IF

INTERACTIVE MODE LOAD &

EXECUTE CYCLIC SCRIPT

(IF TIME TO)

MANAGE AUTOSAVE OR AUTOEXPORT

TAKE USER DEFINED ACTIONS

CONDITIONED BY THE RESULT OF

ALL MEASUREMENTS

SHOW MEASUREMENT

IF DISPLAY MODE

RESULT GOOD?

YES

NO

TAKE USER DEFINED ACTIONS

CONDITIONED BY THE RESULT OF THE SINGLE MEASUREMENT

REPORT, STATISTICS

& BATCH MANAGEMENT

CYCLIC SCRIPT FIRST?

YES

NO

Figure 11

Three different operating modes are outlined here: the DISPLAY mode, theINTERACTIVE mode and the DISPLAYONBAD mode.

If none of these modes are active the QC test proceeds without anymeasurements shown, with simple go-no-go masks, as in Fig.1.

If DISPLAY mode is active then the executed measurements are shown andremain on the screen for a definable amount of time, the test automaticallyproceeds until the end. Fig.2 depicts such a situation.

If INTERACTIVE mode is active the executed measurements are shown and thenthe software prompts for user input . The test sequence is not continued until theuser executes a particular action or actions. It is also possible to loop certainmeasurements for D.U.T. tuning (see Fig.12). Fig.3 depicts such a situation.

If DISPLAYONBAD mode is active then the executed measurements are shownonly if their result is not satisfactory. The sequence is stopped for user acceptance.

Fig.11 shows also the Autosave management which is of great importance forcontrolling the production and for characterizing a batch. This feature is completelyuser definable allowing for binary or text files, operation conditioned by the testresult, coherence with serial number and single test number; the operator can alsobe prompted for file name input.



Two blocks are devoted to the execution of particular actions conditioned by theresult of the single test or the result of all tests. Among these we find:

- messages to the operator- printout of the measurement- execution of custom written software- generation of TTL signals to manage automatic lines- pause for a predefined amount of time- stop the sequence

The third flow diagram, in Fig.12, shows us how the single QC measurement isperformed. As outlined before, CLIO QC relies on the measurements present in thestandard version of the software; the possible measurements within QC are: MLS([MLS]), FFT ([FFT]), Sinusoidal ([SIN]) and Multimeter ([MET]). We will nowcover the keywords which are used to define the tests inside the script.

LOAD LIMITSFILE

LOADREFERENCE

FILE

EXECUTEMEASUREMENT

CHECKRESULT

GOOD

BAD

OPTIONALLOOP

CALCULATEOPTIONAL

PARAMETERS

READY FORNEXT

SET QCOPTIONS

PERFORM MEASUREMENT

INPUT

OUTPUT

Figure 12

To understand this operation we must define two files: the Reference File and theLimits File; these files are the heart of the QC operation, together the QC Scriptthey contribute to define all the parameters of the single measurement.

2.2 THE CYCLIC QC SCRIPT

Again with reference to Fig.11 the last red block, right before the end of the QCtest, represents the QC Cyclic Script execution.

The cyclic script is a particular sequence of QC operations that needs to beexecuted regularly either:- As first action when beginning a QC session - After a certain number of QC tests have been executed

This is useful for testing and re-testing reference quantities that characterize theentire process and maintain traceability to environmental conditions like productiongolden samples.

2.3 THE REFERENCE FILE

The Reference File is a standard CLIO measurement file (extension '.mls', '.fft','.sin', or '.met') created within its relative menu; it contains most of the settings



needed to fully configure your measurement. Just as CLIO resets the measurementcontrol panel to the settings of the file loaded from disk, the QC processor does thesame job; in this easy but effective way of operating you will be sure that, forexample, the sampling frequency of your QC MLS measure will be the one youchose, or the display settings will be the same as when you saved the referencefile. And all this is defined, inside the QC script, with a single text line:

REFERENCE=myreferencefile.mls

where we imagined that you gave the name 'myreferencefile' to a saved MLSmeasurement.

One very important setting stored within the reference file is if the measurement ismono (only channel A acquired) or stereo (channel A and B acquiredsimultaneously).

2.4 THE LIMITS FILE

The Limits File is a text file ('.lim' extension) defining the frequency mask orquantities needed to check the executed measurement. The syntax used is thesame as the QC script. A Limits file can be as simple as:

[UPPER LIMIT DATA]100 +5500 +35000 +110000 +5[LOWER LIMIT DATA]100 -5500 -35000 -110000 -5

In principle nothing else is needed to define the basic measurement; here is anexample of a section of a QC script defining a MLS measurement:

[MLS]REFERENCE=MYREFERENCEFILE.MLSLIMITS=MYLIMITSFILE.LIM

An interesting keyword to add is COMMENT that let’s you give a brief description ofthe QC test that will be output during the measurement and inside reports:

[MLS]COMMENT=FREQUENCY RESPONSEREFERENCE=MYREFERENCEFILE.MLSLIMITS=MYLIMITSFILE.LIM

While performing a QC measurement CLIO can calculate more parameters from thedata acquired and have these parameters to concur with the final result. As anexample it is possible to make a polarity check within a MLS frequency responsemeasurement or make a T&S parameters check within an impedance measurement.The following script adds the polarity check to the former MLS test.

[MLS]REFERENCE=MYREFERENCEFILE.MLSLIMITS=MYLIMITSFILE.LIM



POLARITY=1

Here is a list of the parameters that can be calculated within each measurement:

Sinusoidal - Frequency response and impedance response (mono or stereo tests)- Average (or single frequency) level- Sensitivity (average or up to eight frequencies)- Polarity- Total harmonic distortion response- Single harmonic response (from 2nd to 10th)- Fast-Track Rub&Buzz response- T&S parameters (Fs,Qt,Qe,Qm,Cms,Mms,Mmd,Vas,Bl,dBSPL,ZMin)- Loudness Rating (RLR, SLR, STMR)

MLS&CHIRP- Frequency response or impedance response (mono tests) - Average (or single frequency) level- Sensitivity (average or up to eight frequencies) - Polarity- T&S parameters (Fs,Qt,Qe,Qm,Cms,Mms,Mmd,Vas,Bl,dBSPL,ZMin)- Loudness Rating (RLR, SLR, STMR)

FFT - Frequency response with definable stimulus (mono tests, alsointeractive) - Average (or single frequency) level- Sensitivity (average or up to eight frequencies)

METER - SPL, Volts, THD, IMD single parameter (mono tests, also interactive)

When you have taken a single channel mono measurement you define only onelimits file.

2.5 LIMITS FILE FOR TWO CHANNELS STEREO OPERATION

When you have taken a simultaneous two channels stereo measurement you maydefine the following limits files:

A) One single Limits file which is valid and shared for both channels; this is the casewhen both measurements refer to the same unit like the two channels frequencyresponse of a headphone or of a stereo equipment. A stereo sinusoidal test may bedefined as:

[SIN]REFERENCE=MYREFERENCEFILE.SINLIMITS=MYSTEREOLIMITSFILE.LIM

B) Two different Limits files one per measured channel; this is the case when thetwo measurements refer to two different quantities like a frequency responsetogether an impedance response. The LIMITS keywork, in this case, is substitutedby the two keywords LIMITSA and LIMITSB. A stereo sinusoidal test may bedefined as:

[SIN]REFERENCE=MYREFERENCEFILE.SINLIMITSA=MYRESPONSELIMITSFILE.LIMLIMITSB=MYIMPEDANCELIMITSFILE.LIM



3 THE CLIO QC TCP/IP MEASUREMENT SERVER

This is the CLIO answer to the general request of being able to control and use QCfeatures inside custom applications.

It is an imperative need when audio testing is a part of a more complex QC process(like in a cell phone QC test procedure when you must test also the display andother parts).

The choice of TCP/IP approach presents several advantages:

1) No additional learning curve; the same CLIO QC script commands areused

2) Prevents the engineer to deal with complex API programming

3) It is independent from the Operating System, Programming Languageand kind of PC.

4) It can be run locally or from another network connected PC.

5) It is possible to write applications that control more than one QC testworkstation.

3.1 INVOKING THE QC SERVER

To invoke the CLIO quality control server simply run CLIO passing it the “TCP”parameter. You may define a shortcut with the following target program:

“C:\Program Files\Audiomatica\CLIO11\Clio.exe TCP”

CLIO will run and start listening on the port defined in the CLIO Options>QCsettings dialog (see chapter 19) being port 1234 the default one.

The CLIO desktop will also show this particular operating condition in the maintoolbar:

From this moment it is possible to connect to CLIO and receive the variousmeasurements services that it is capable of.



3.2 CONNECTING TO THE QC SERVER

It is possible to connect to the CLIO QC server with any custom written clientapplication that opens a TCP socket (we will see an example later) or with astandard telnet application (like Microsoft Telnet).

The connections parameters are:

hostname Network name of PC or ‘localhost’ for same PC

port CLIO TCP port (default 1234)

Let’s see how to connect a telnet client application (we will use CRT 3.4) run in thesame computer where CLIO resides.

As soon as the connection is invoked the CLIO QC server will answer with thewelcome greeting:

The connection is established! QC services are ready for you.



3.3 INTERACTING WITH THE QC SERVER

Your client application interacts with CLIO sending the standard ASCII scriptcommands; CLIO executes the commands and sends back the result of themeasurements.

Let’s now execute a simple MLS measurement. We will use the same exampledescribed later (My First QC Script). The syntax is identical:

[MLS]OUT=1.000 VINA=0INB=0REFERENCE=LOOP.MLSLIMITS=LOOPMLS.LIM

If we send these commands to CLIO we get the following:

You can see how the data exchange takes place. After each line of command is sentthe server sends back an acknowledgment stating that the command has beenreceived and that it is OK. At this time the sequence has not been closed yet andthe measurement has not been done. The server needs to know that the sequenceof commands that defines the measurement has ended; there is the special executecommand [] (two empty brackets) that is needed, at the end, to tell CLIO toexecute the measurement.

After we give the execute command ([]) the measurement starts and the result isfed back to our application. The first line of the result is the global test result whileeach subsequent line details all the single checks that have been done and that



participate to the global result.

To see more tests in action we may add a level check and a polarity test. To do thiswe must add the following to the limits file ‘loopmls.lim’:

[LEVEL]UPPER=2LOWER=-2

And we must add the following to the commands sent:

POLARITY=1

We get the following situation:

You notice now that the result is detailing all the three checks that the MLSmeasurement has done (response, global level and polarity).

The example details how to execute a measurement; single commands can also besent that perform all standard operations. To close the channel A in-out loop simplysend CLIO the following:

[SETLOOPA][]

In the above example CLIO is behaving as a server and is visible on the WindowsDesktop. It is possible to hide CLIO from end user sending the command:



[HIDECLIO][]

CLIO will disappear and remain minimized in the Windows application bar; to seeCLIO again send:

[SHOWCLIO][]

3.4 NOTES ABOUT QC SERVICES

The Quality Control operation when requesting TCP services differs from the normalcondition when the QC Script processor is active.

In this case many tasks are handled by the client application that is requesting theservices and are not performed by CLIO; for example there is no serial numbermanagement.

The main difference is that no QC test, formed by various single measurements, isdefined and managed by CLIO like in a QC script; the TCP server can be configuredand then executes endlessly all the commands and measurements it is requested todo; it has no knowledge of how many single measurements form a complete QCtest.

TCP Operation and Server messages

When dealing with a network service like the CLIO TCP server the client applicationreceives back answers for each text command sent.

We find the following server responses:

200 Start Command OKUsually given when a bracketed keyword is sent

200 Additional Command OKUsually given when a keyword defining a section is sent

400 Unknown Command

400 Unknown Additional Command

200 OK Given when a command (not a measurement) is executed

200 GOODGlobal result given at the end of a measurement

200 BAD Global result given at the end of a measurement

200 GOOD Response, 200 GOOD Polarity etc. etc.Single results given at the end of a measurement

Note the particular syntax of these answers. They are all initiated by a number thatis related to network operation and gives information about the correct interactionbetween client and server. We find:

200 Correct



400 Usually an error is occurred

Autosaving

During TCP operation the QC single test numbering is disabled and does not takeplace in defining the name of the autosaved data file (see later). If autosaving isactive CLIO will give the following names to files:'tcpresponse.txt' measurements exported in ASCII 'tcpresponse.mls' MLS measurements'tcpresponse.sin' Sinusoidal measurements'tcpresponse.fft' FFT measurements 'tcpresponse.met' Multimeter measurements

Please note also the following differences with standard QC operation:

- No serial number management is performed

- No batch management is performed

- No production report files are saved

- No statistical information are calculated



4 HANDS ON QC

4.1 WHAT TO KNOW ABOUT QC SCRIPTS

A quality control script is a text file that stores information in logical groupings,called sections.Each section is initiated by a bracketed keyword in the form [keyword].Within each section, QC definitions are stored in named keys.Keys within a section take the form keyword=value.

For example the section called [GLOBALS] defines several settings useful all alongthe test sequence:

[GLOBALS]COMPANY=MY COMPANYTITLE=MY QUALITY CONTROLBATCH=MY PRODUCTION BATCH NAME

It is possible to input comment lines initiated by a semicolon. It is not possible tostart a comment after a keyword.

;this is a correct comment lineCOMPANY=MY COMPANY ;this comment is not allowed

With an understanding of these brief notes you are ready to write a QC script.

4.2 HOW TO WRITE MY FIRST QC SCRIPT

You may write your script with any text editor that stores plain ASCII files (usually'.txt' ones), like Notepad; the only thing you should remember is that QC scriptsmust have the '.qc' extension while limits files use the '.lim' extension; the commonbehavior of Windows to hide registered file extensions sometimes renders thisaction difficult. It is not uncommon to believe you have saved a file with, say, thename 'myfile.qc' (where you tried to force the extension) and then find it actuallysaved as 'myfile.qc.txt' because the text editor automatically appended theregistered extension.

You may write your script directly by editing it within the QC control panel textdisplay; in this case the extension management is guaranteed by CLIO and you willbe able to use some tools, like measurements capture, that are of help duringeveryday jobs. By doing it like this it is possible to immediately test the script bypressing Go.

Let's now write our first QC script.

Have your CLIO system in the same setup as when you performed the systemcalibration: output A connected to input A; see chapter 3 for details. Don't connectany external device to the system. Set output level at 0dBu and input sensitivity at0dBV (see Chapter 4 for details). Have the default settings loaded.

Open MLS; press Go. You should obtain a straight line as in Fig.23. Expand thedisplay to obtain 2dB/div ans set upper Y scale value to 4dBV. Save thismeasurement as 'Loop.mls'.



Figure 23

Now open the QC control panel. Press N, we are starting a new script. Press Ctrl-Eto exit edit mode and then press L to enter Limits Text mode. Input the followingfrequency masks as limits:

[UPPER LIMIT DATA]20 1.230 0.715000 0.720000 1.2[LOWER LIMIT DATA]20 -1.230 -0.715000 -0.720000 -1.2

Press F2 and save the limits file as 'loopmls.lim'. Now click now on the (script)button and then click on the (capture) button. Your blank text display shouldnow be filled with your first QC script:

[MLS]OUT=1.000 VINA=0INB=0REFERENCE=LOOP.MLSLIMITS=LOOPMLS.LIM

It is a good practice to add the following comment line:

COMMENT=FREQUENCY RESPONSE

Click on the go button; the QC processor should execute a QC test performingan MLS measurement, displaying it together with the defined limits, everything asin Fig.24; the text display should now present information on the executed test.



Figure 24

Let's now complete this first exercise by adding a Multimeter measurement of leveland total harmonic distortion at 1kHz.

Press F4 to open (and run) the Multimeter control panel, then click on the generator button to switch the generator on and play the default 1kHz sinusoid.Now press T to stop measuring; save this measurement as 'loop.met'; Fig.25should be what you have in front of you.

Figure 25

Now press Ctrl-Q and then L to go back to inputting a limits file definition. Inputthe following:



[UPPER LIMIT DATA]VOLTAGE=1.1THD=0.01[LOWER LIMIT DATA]VOLTAGE=0.9THD=0.0001

Save this as 'loopmet.lim'. Now click on the button and position the cursor insidethe text display after the last line of text; as before, click on the capture button andthe following lines should be added and you are ready for this new QC test.

[MET]OUT=1.000 VINA=0INB=0REFERENCE=LOOP.METLIMITS=LOOPMET.LIM

It is a good practice to add the following comment line:

COMMENT=LEVEL+THD

Now pressing the Go inside QC executes this two-measurement QC test sequence;Fig.26 shows the test at its end.

Figure 26

This concludes our first approach to QC script writing and debugging. All the filesnecessary to "study this lesson" are downloadable from Audiomatica website.

The 'loop.qc' script is doing exactly what has just been described with a difference:measurements are performed in interactive mode; just load it and run it to feelthe differences.



5 RUNNING A QC SCRIPT

The QC environment can be programmed to act in several ways while presentingthe operator different interactive panels to customize her/his user experience.

5.1 DESKTOP AND WINDOWS MANAGEMENT

The CLIO desktop, as default behavior, automatically handles the measurementwindows needed for the QC test:

1) Closes unnecessary open windows

2) Maximizes the graphical display of each window

3) Tiles the open windows to fill the desktop

It is possible to disable the automatic management of the windows using theTILEWINDOWS keyword. Add:

[GLOBALS]TILEWINDOWS=0

Then resize the measurement windows as you like. Their relative positions andsizes will not be changed until next software run.



5.2 THE QC RESULT PANEL

The QC Result panel usually accompanies QC sessions where measurement displayis not needed. This results in a situation with simple go-no-go masks for use withcompletely automatic lines or for operators who don't need to take particularactions with respect to the test result.

To activated the QC Result panel from within the QC script use the DISPLAY=0keyword.

Note: for maximum QC test speed use the QC Result display and don'tshow single measurements as the display of graphical objects andmeasurement curves may employ a lot of processor time.

The QC Result panel can be forced to appear at the end of the QC sequencepressing the button.



If Shrink QC result is selected in the associated drop down menu the QC resultpanel will appear in a minimized version.

5.3 THE QC BANNER

The QC Banner is managing information and messages given to the operator whilein Interactive mode.

5.4 THE QC REPORT PANEL

The QC Report panel serves as an interactive tool which is of great help forinspecting a production while it is tested; it is composed by two tree views namedSTATISTICS and TEST REPORT these handle all the information pertaining toyour QC session in a very compact form.

The QC Report panel can be kept open during the tests and it accompanies thework in a really effective visual form.

Under STATISTICS you find information about:- QC test and Company names



- Date of the first unit tested- Name of the production batch- First serial number tested- Total number of units tested, number of “good” and “bad” units

Under TEST REPORT you find information about:- DUT test result with serial number and time of production- Single tests results- Names of the saved files

The QC Report panel is also the starting point for reviewing a saved measurementas described below. The name of the saved file is a sensible area where you candouble-click to review the measurement.

5.5 REVIEWING A MEASUREMENT

During a QC tests session it is possible to review a measurement that has beensaved to disk. This is important when, for example, trying to understand why ameasurement went bad. As we saw before the QC report panel indicates all thenames of the files that have been created during the test execution, under therelative serial number and single test number.

As soon as a QC sequence is terminated simply open the tree view of your interest,identify the measurement you want to inspect and double click on its name. CLIOloads the measurement as if it were performed inside the running QC, together withits pertinent limits and executes all the calculations defined in the QC script endingwith the result check and display. The following diagram describes such a process;compare it with Fig.12.



LOAD LIMITSFILE

LOADREFERENCE

FILE

LOADMEASUREMENT

FROM DISK

CHECKRESULT

GOOD

BAD

CALCULATEOPTIONAL

PARAMETERS

DISPLAYRESULT

DOUBLE-CLIKON REPORTTREE VIEW

REVIEW MEASUREMENT

Reviewing a saved measurement from within QC is different from simply openingthe file from the measurement control panel; in this second case no post processingdue to QC operation is applied. Figure below shows a measurement (black curve)reviewed inside QC with its limits (red and blue curves) and the same measurementloaded from the measurement control panel (purple curve); the shift in level is dueto QC operation when it separately checks for relative level and frequency behavior.

Note: the review operation can be done only when inside a QC session; if CLIO isexited, then later QC is started again a new QC session will be created; reportinformation and review operation will only apply to the new session.

5.6 THE SKIP LAST BUTTON

When a QC test is finished it is possible to null its result by pressing the Skip Last button. All information saved with the test will be erased comprising serial

number increment and statistical data. The production report will mark the unit as'SKIPPED'.



6 NOTES ON LIMITS CURVES

As previously outlined the QC processor needs limits data in order to perform therequired checks. This data is saved within the limits files and usually represent afrequency mask (for frequency response and impedance tests) but they can alsodefine a single value check (like, for example, a Qms test).

When dealing with frequency checks the options defined affect the way thefrequency masks are calculated, the way data is displayed on screen and the waythat the result is checked. It is also possible to add an average or single frequencylevel check that concurs with the final result.

Fig.27 shows us the procedure for calculating the frequency mask after the limitsfile is loaded into memory. You can see that the frequency data sets saved under[UPPER LIMIT DATA] and [LOWER LIMIT DATA] are treated differently if the limitsare absolute or relative or if an aligned point is defined (see later).

YES

GET ALIGNPOINTDATA

RELATIVE?GET

REFERENCEDATA

ALIGNED?CALCULATE

LIMITSCURVES

NO

YES

NO

LOAD LIMITSFILE

Figure 27

Fig.28 shows us the way a frequency check is performed and the measurement ispresented on screen. You may appreciate the presence of an average level (orsensitivity) check or a single point (aligned) level check that concurs with the finalresult. When a level (or sensitivity) check is defined, either the measuredcurve or the limits curves are shifted if presented on screen; in this way it ispossible to appreciate the frequency behaviour of the measured curve without theeffect of a difference in sensitivity which is checked separately.

CHECKLEVEL

AVERAGELEVEL CHECK?

YESSHIFT LIMITS

CURVESALIGNED?

CALCULATELEVEL @

ALIGN POINT

CALCULATELEVEL IN

LIMITS BAND

FLOATINGLIMITS?

SHIFTMEASUREMENT

CURVE

CHECKRESPONSE

NO

YES

NO

NO

YES

FINALRESULT

Figure 28a frequency plus average level check



CHECKSENSITIVITY

SENSITIVITYCHECK?

YESSHIFT LIMITS

CURVES

SINGLEFREQUENCIES?

CALCULATESENSITIVITY @

DEFINEDFREQUENCIES

CALCULATESENSITIVITY INLIMITS BAND

FLOATINGLIMITS?

SHIFTMEASUREMENT

CURVE

CHECKRESPONSE

NO

YES

NO

NO

YES

FINALRESULT

Figure 28b frequency plus sensitivity check

As a final, but not less important note, we show an alternative method to define alimits file; it is possible to input the frequency mask as a text file as below.

[UPPER LIMIT DATA]FILE=UPPER.TXT[LOWER LIMIT DATA]FILE=LOWER.TXT

The files 'upper.txt' and 'lower.txt' are export ASCII files that may be produced byother applications or CLIO itself.

The 'upper.txt' file may look like:

Freq[Hz] dBV100 5500 35000 110000 5



6.1 ABSOLUTE VS. RELATIVE FREQUENCY LIMITS

The following limits file defines an absolute frequency limit.

[ABSOLUTE][UPPER LIMIT DATA]200 100300 9710000 9715000 100[LOWER LIMIT DATA]200 82300 8510000 8515000 82

The frequency mask is shown in the first image below.

The following limits file defines a relative frequency limit.

[RELATIVE][UPPER LIMIT DATA]200 5300 210000 215000 5[LOWER LIMIT DATA]200 -5300 -210000 -215000 -5

The frequency mask is shown in the second image above. Relative means withrespect to the reference file defined in the QC test. Data values will be addedand subtracted to the reference value at the specified frequencies.

Relative data values may be considered as percentages. The following keywords isrequired.

[RELATIVE]PERCENT=1

The above mask may be defined for an impedance measurement curve andconsidered as percentage; in this assumption it the calculated limits curves would



differ by ±2% in the 300-10000Hz region while ±5% outside with respect to thereference.

An important feature for a relative file is the possibility of adding a frequencyjitter to the calculated limits curves. This quantity is expressed in fractions ofoctaves and tells how much jittering is applied to the limits. The effect, shown inthe below curves is to allow rapidly changing (but small) frequency behaviors of themeasured curves while not loosening too much the mask.

The limit curves in the left figures have no jitter but may be problematic during QCoperation, easily giving false negatives, due to the break-up effects in the higherpart of the spectrum.

Adding a 1/3 of octave jittering with:

[RELATIVE]FREQJITTER=0.3

You obtain the relative limits as in the right figure which cure the problem notgiving rise to false negatives while keeping the mask tight.

It is possible to input up to 2048 frequency points to define the check mask. TheQC processor will execute the check starting from the first frequency point, endingat the last; no check will be done outside this frequency range.

Inside a frequency limits file it is possible also to define frequency masks forexecuting a QC check on the following:- Average (or single frequency) level- Sensitivity (average or up to eight frequencies)- A/B difference between channels in a stereo measurement- Sinusoidal THD, Single harmonic or Fast-Track Rub&Buzz response - T&S parameters (Fs,Qt,Qe,Qm,Cms,Mms,Mmd,Vas,Bl,dBSPL,ZMin)- Loudness Rating (RLR, SLR, STMR)

A frequency limit file can be applied to an MLS, Sinusoidal or FFT test. To define alimits file for a Multimeter measurement see later.



6.2 AVERAGE LEVEL CHECK

The following limits file defines an average level check inside the same relativefrequency limit shown before.

[RELATIVE][LEVEL]UPPER=3LOWER=-3FREQHI=5000FREQLO=400[UPPER LIMIT DATA]200 5300 210000 215000 5[LOWER LIMIT DATA]200 -5300 -210000 -215000 -5

When a level check is defined inside a limits file the QC result is actually acombination of two separate checks; one is the frequency behavior of themeasurement compared against the frequency mask, the second is a level checkwhich compares the average level of the measured curve with the average level ofthe reference.

The average level is calculated within the frequency extremes defined by FREQHIand FREQLO as shown in figure.

As default, if FREQHI and FREQLO are not defined, the levels are calculatedaveraging in the frequency band defined by the extremes frequencies ofthe limits.

The next figure shows such a situation; the title of the measurement control panelreports the level check.



The level check shown means that the value of the measurement averaged in theband shown is 0.09dB higher than the reference average level in the samefrequency band.

The measured curve is shifted from this value and then the frequencycheck is performed.

The level shift means that the curve is displayed with a different level fromthe measured one.

As two separate checks are done there may be two distinct cases when a unitresults in a bad report. The following figures try to explain these two cases.

Figure shows us the case of a unit is testing bad because the frequency behavior isnot good while the average level is OK.



The last figure, instead, shows us the case of a unit is testing bad because theaverage level is not good while the frequency behavior is OK.



6.3 ALIGNED MASK

The following limits file defines a single point level check with a frequencymask aligned to it.

[ABSOLUTE][LEVEL]UPPER=3LOWER=-3ALIGNFREQ=5000ALIGNLEV=90[UPPER LIMIT DATA]200 5300 2800 21000 63000 64000 27000 215000 8[LOWER LIMIT DATA]200 -5300 -210000 -215000 -5

The align point (in the example 90dBSPL@5000Hz) is used to build the frequencymask (that is specified relative to it) and also to identify the frequency at which toperform the level check.

Figure shows a mask aligned to the point (90dBSPL@5000Hz). The level checkmeans that the value of the measurement at 5000Hz is 0.22dB higher than thealign point.

The measured curve is shifted from this value to pass at exactly 90dBSPLat 5000Hz; then the frequency check is performed.

The level shift means that the curve is displayed with a level different fromthe measured one.



6.4 SENSITIVITY CHECK

The following limits file defines a sensitivity check inside a relative frequencylimit.

[RELATIVE][SENSITIVITY]UPPER=102LOWER=100[UPPER LIMIT DATA]200 10500 101000 51500 52000 104000 10[LOWER LIMIT DATA]200 -10500 -101000 -51500 -52000 -104000 -10

As per the average level check, when a sensitivity check is defined inside a limitsfile the QC result is actually a combination of two separate checks; one is thefrequency behavior of the measurement compared against the frequency mask, thesecond is a sensitivity check which compares the sensitivity of the measured curvewith the defined upper and lower limits.

It is possible to calculate sensitivity at discrete frequencies (up to eight) andaverage them together.

[SENSITIVITY]FREQ1=500FREQ2=1000FREQ3=2000UPPER=102LOWER=100



6.5 FLOATING LIMITS VS. FLOATING CURVES

When an average or single frequency level check is defined it is possible to definefloating limits instead of floating curves using the [FLOATING] keyword.

[RELATIVE][FLOATING][LEVEL]UPPER=3LOWER=-3[UPPER LIMIT DATA]200 5300 210000 215000 5[LOWER LIMIT DATA]200 -5300 -210000 -215000 -5

In this case the measured curve is presented on screen with correct valueswhile the limits curves are moved around it.



6.6 SINUSOIDAL A/B STEREO DIFFERENCE CHECK

When executing a stereo sinusoidal frequency response measurement it is possibleto activate quality control checks over the calculated difference between the twochannels.

The display is possible only for one curve chosen among the pool of the curvescalculated within a single sinusoidal test.

Note: When a distortion curve is displayed, its graphical properties are definedwithin CLIO Otpions>Graphics>” QC Curve C” .

For A/B stereo difference QC check do the following:1) Execute and save a stereo reference measurement.2) Define a limits file adding the limit definition:

[A/B UPPER LIMIT DATA][A/B LOWER LIMIT DATA]

Select A/B calculated curve for display:

[A/B DISPLAY]



6.7 SINUSOIDAL THD AND FAST-TRACK RUB&BUZZ CHECK

When executing sinusoidal frequency response measurements it is possible toactivate quality control checks over calculated THD, Rub-&Buzz or single harmonic(from 2nd to 10th) response curves.

Calculation and QC check is possible for any distortion curve.The display is possible only for one curve chosen among the pool of the curvescalculated within a single sinusoidal test.

Note: When a distortion curve is displayed, its graphical properties are definedwithin CLIO Options>Graphics>” QC Curve C” .

For THD and Harmonics QC check do the following:1) Execute and save a reference measurement with “THD Enabled” under settings.2) Define a limits file adding the limit definition:

[THD UPPER LIMIT DATA]

for THD and for any harmonic (if desired):

[2 UPPER LIMIT DATA][3 UPPER LIMIT DATA]....[10 UPPER LIMIT DATA]

Select one calculated curve for display:

[THD DISPLAY]

For Fast-Track Rub&Buzz QC check do the following:1) Execute and save a reference measurement with “R&B Enabled” under settings.2) Define a limits file adding the limit definition:

[RUB+BUZZ UPPER LIMIT DATA]

Select rub&buzz curve for display:

[RUB+BUZZ DISPLAY]

NOTE 1: If more than one curve is selected for display only one will be displayed,the others only calculated and QC check done; to inspect the curves not diplayedafter a QC test is finished you must release the measurement and operate theproper buttons within the sinusoidal menu.

NOTE 2: If a level or sensitivity check is performed within the QC check and thedistortion data are expressed in dB units (not % units) the calculated limit masks(R&B, THD and nth Harmonic) will be shifted to take into account the sensitivitydifference with the reference.



6.8 THIELE&SMALL PARAMETERS CHECK

It is possible to execute QC tests of the following T&S parameters:

Qt, Qe, Qm, Fs, Cms, Mms, Mmd, Bl, Vas, dBSPL and ZMin.

To evaluate the first four parameters it is necessary to input the value of the DCresistance of the voice coil with the keyword REDC.To evaluate the remaining parameters, by means of a simplified estimation routine,it is necessary to input the value of the driver diameter with the keywordDIAMETER and one of the following fixed quantities: KNOWNMMD (fixed mass) orKNOWNMMS (fixed mass plus air load) or KNOWNCMS (fixed compliance).

The following limits file defines a T&S parameters check inside a limits file with afrequency mask for an impedance response. The parameters checked are Qt, Qe,Qm and Fs.

[TSPARAMETERS]QTUPPER=0.3QTLOWER=0.05QEUPPER=0.3QELOWER=0.05QMUPPER=5QMLOWER=2FSUPPER=90FSLOWER=50REDC=5.5[UPPER LIMIT DATA]29.89 142.3540.52 161.19102.15 161.19152.62 143.53[LOWER LIMIT DATA]29.89 11.2949.23 20.0064.33 45.8876.28 47.0698.49 22.35141.87 11.7The following section defines a T&S check of Qts, Fs, Cms, Bl and ZMin havingfixed the mechanical mass Mmd value.

[TSPARAMETERS]REDC=6.2DIAMETER=110KNOWNMMD=10.7952QTSUPPER=0.6QTSLOWER=0.3FSUPPER=90FSLOWER=50CMSUPPER=1.1CMSLOWER=0.8BLUPPER=6.5BLLOWER=6ZMINUPPER=7.5ZMINLOWER=7



6.9 LOUDNESS RATING CALCULATION AND CHECK

It is possible to execute QC tests of the following loudness rating indicators:

RLR, SLR, STMR.

The following limits file defines a loudness rating parameters check inside a limitsfile with a frequency mask for an frequency response.

[LR]SLRUPPER=11SLRLOWER=5[UPPER LIMIT DATA]100 3200 1.53000 1.55000 3[LOWER LIMIT DATA]100 -3200 -1.53000 -1.55000 -3



6.10 MULTIMETER LIMITS FILES

The following limits file defines a multimeter QC check.

[UPPER LIMIT DATA]VOLTAGE=0.78THD=0.01[LOWER LIMIT DATA]VOLTAGE=0.77THD=0.0001

The parameters available are:

- PRESSURE- VOLTAGE- FREQUENCY- THD- IMD



7 MANAGING PRODUCTION BATCHES

Managing a production batch is a rather complex while delicate topic as it involvesdiverse needs of diverse areas inside your company.

CLIO QC handles your batch doing the following:

- Maintains a directory structure where different files are saved- Automatically saves production report files- If requested autosaves data files- Handles 24 characters alphanumeric serial numbers- Auto increments serial number and maintains its coherence- Calculates statistical data about the batch

The result is that you will find the production well documented both for yourinternal purposes aimed to achieve the highest quality standard and also forinterfacing with your client who requests technical information about the units.

7.1 DIRECTORIES CREATED BY CLIO QC

Suppose you saved your script inside the directory 'My qc'. When you run the scriptCLIO automatically creates one or more directories under 'My qc'. There are fourcases depending on the option you set:

1) No Autosave is active. A Batch is not defined.CLIO creates the 'Report' directory where all the production report files aresaved. Fig.37 shows this situation.

2) Autosave is active. A SaveFolder is not defined. A Batch is not defined.CLIO creates the 'Report' directory where all the production report files

are saved. It also creates the 'Autosave' directory where all data files are saved.Fig.38 shows this situation.

Figure 37 and 38

3) A Batch is defined and is named 'My Batch'. A SaveFolder is not defined.CLIO creates the 'My Batch' directory where all the production report and alsodata files are saved. Fig.39 shows this situation.

4) A SaveFolder is defined and is named 'My Savefolder'.CLIO creates the 'My Savefolder' directory where all the production report andalso data files are saved. Fig.40 shows this situation.

Figure 39 and 40



7.2 PRODUCTION REPORT FILES

Suppose that today, June 6, 2002, at 6:46, you started a production of yourdevices; the batch, named 'My Batch', ended yesterday with unit number 100.

After two units tested CLIO will add, under the folder 'My Batch', the followingreport files:

'production_06-06-02_6.46.19.txt''101.txt''102.txt'

After 20 units tested:

'production_06-06-02_6.46.19.txt''101.txt''102.txt'...........'120.txt'

If you stop the production, exit CLIO, and then restart it at 7:01, after two moreunits tested:

'production_06-06-02_6.46.19.txt''production_06-06-02_7.01.05.txt''101.txt''102.txt'...........'122.txt'

The files 'production_date time.txt' describe the QC session. They look like:

STATISTICSMY COMPANYMY QUALITY CONTROLBATCH = My BatchDATE = 06-06-02INITIAL SN = 101TOTAL TESTS = 2GOOD = 2BAD = 0

TEST REPORTUNIT N.102 GOOD 6.46.24

1 GOOD MLSResponse GOODC:\Program files\Audiomatica\CLIOpci\Data\My qc\My

Batch\102_1.mls2 GOOD MET

Voltage:0.775Vrms GOODTHD:0.006% GOODC:\Program files\Audiomatica\CLIOpci\Data\My qc\My

Batch\102_2.metUNIT N.101 GOOD 6.46.19

1 GOOD MLSResponse GOODC:\Program files\Audiomatica\CLIOpci\Data\My qc\My

Batch\101_1.mls



2 GOOD METVoltage:0.775Vrms GOODTHD:0.006% GOODC:\Program files\Audiomatica\CLIOpci\Data\My qc\My

Batch\101_2.met

The files 'serialnumber.txt' describes the single QC test and look like this:

1 GOOD MLS Response GOOD2 GOOD MET Voltage:0.775Vrms GOOD THD:0.006% GOOD06-06-02 6.46.24UNIT N. 102 GOOD

7.3 AUTOSAVED DATA FILES

Again supposing we are in the situation of the preceding paragraph let's see howdata files are saved. As it can be seen from the report files our QC test consists of aMLS and a Multimeter measurement. As the MLS test is defined before theMultimeter inside the script then it assumes number 1 as single QC test while theMultimeter test assumes number 2; this is already clear from the report files above.

After two units tested we find the following measurement files:

'101_1.mls''101_2.met''102_1.mls''102_2.met'

As you see the QC single test numbering is integral part of the name of theautosaved data file.

7.4 STATISTICAL INFORMATION ON MEASURED DATA

Statistical information characterizing the production can be obtained by CLIO usingthe STATISTICS keyword under [GLOBALS].

CLIO will save, under the report directory, the following files:

- One file named 'data_table.txt' with statistical information on all the measuredparameters.

- One file named 'avg_testnumber.txt' for each response test defined containingthe average response for that test.

- One file named 'sdmax_testnumber.txt' for each response test definedcontaining the average response plus twice the standard deviation for that test.

- One file named 'sdmin_testnumber.txt' for each response test definedcontaining the average response minus twice the standard deviation for that test.



The statistical files keep track of the all the units saved within a batcheven if the production is stopped and then restarted.

Let's now see what the 'data_table.txt' looks like; supposing the same case of19.7.2, after two tests, we would have the following:

SN Voltage THD 101 0.775 0.006 102 0.775 0.006

Avg 0.775 0.006SDMax 0.776 0.006SDMin 0.775 0.006

The other response files representing average and standard deviation curves maybe imported within each control panel with the Import feature recallable with Shift-F3.

7.5 SERIAL NUMBER MANAGEMENT

There are several ways to handle the serial numbers of your devices and tomaintain their coherence through all the production of one batch.

Two different strategies are possible with respect to serial number management:1 - CLIO handles and manages an 8-digit numeric serial number. This is thedefault operation.2 - CLIO accepts a 30 characters alphanumeric serial number; its managementis left to the user.

To activate the second option use the AUTOSN=0 keyword (default is AUTOSN=1).

[GLOBALS]...AUTOSN=0......[SNINPUT]

The operator is prompted for serial number input using the [SNINPUT] keyword.Input can be done with any kind of bar code reader.

It is also possible to manually input the serial number before starting the test; todo this just click on the button.

Under default operation (AUTOSN=1) the 8-digit serial number is automaticallyincreased after the end of the test. It is possible to avoid a bad unit increasing theserial number using the INCREASEONBAD=0 keyword.

Set INCREASEONBAD=0 if you want only good units to have a serialnumber, report , statistical and autosave management; this works alsowhen AUTOSN=0.

The operator, under her or his judgment, can force the final result of a bad test ifthe keyword PROMPTFORGOOD=1 is used.



8 INTERACTING WITH EXTERNAL HARDWARE

The interaction with external hardware gives CLIO the possibility of realizing semior fully automatic production line QC tests. Several keywords have been introducedto implement this functionality (see to reference section for a complete listing).

8.1 INPUT SENSITIVITY AND OUTPUT VOLTAGE CONTROL

As we have already seen it is of fundamental importance to correctly set CLIO'sinput sensitivity and output level. The IN and OUT keywords are used for this. Thescript below sets the input sensitivity at 10dBV and output level at 0dBu. Thesenumbers also directly appear also in the main tool bar of CLIO.

...IN=10OUT=0...

The OUTUNITS keyword can be used, under [GLOBALS], to define the output levelunit of measure; you may choose either V, dBV or dBu; default is dBu. To output 1Vsimply write:

[GLOBALS]OUTUNITS=V...OUT=1...

or, even simpler,

...OUT=1V...

If you feed the output to a power amplifier the resulting signal at amplifierterminals will be amplified by the gain of the amplifier. It is possible to take thiseffect into account and specify the output level directly at the amplifier’s output inthe particular case you are using a CLIOQC Amplifier & SwitchBox. The followingscript can be used to set 2.83V at the output of the amplifier.

...OUTQCBOX=2.83V...

8.2 QCBOX MODEL 5 DC OUTPUT CONTROL

The QCBox Model 5 Amplifier&SwitchBox has the capability of superimposing a DCvoltage to the generated signal. It is possible to manage this DC voltage with theQCBOXDCOUT keyword. This DC voltage ranges from -20 to 20V. The scriptbelow sets a 2V DC at speakers terminals.

[PERFORM]QCBOXDCOUT=2



8.3 CLIOQC AMPLIFIER&SWITCHBOX CONTROL

Custom keywords have been implemented to easily control all the internal functionsof this unit:

[SETINPUT1] Selects input 1 of the CLIOQC Amplifier & SwitchBox.








[SETIMPEDANCE]Selects impedance mode of the CLIOQC Amplifier &SwitchBox.

[SETISENSE] Selects I Sense mode of the CLIOQC Ampli&SwitchBox.

Specific keywords are dedicated to the QCBox Model 5; these keywords have noeffect in the case of earlier versions of the unit.

QCBOXCURRENTLIMITCurrent limit (A) for Model 5 operation. Ranges from 0 to 10.

QCBOXDCOUTDC voltage (V) to be output by Model 5 superimposed to generatedsignal. Ranges from -20 to 20.

QCBOXINITIALBYTE8-BIT binary value that will be output from Model 5 port at startupbefore QC script execution.

QCBOXOUTBIT0Status (=1 or =0) of the bit that will be output from Model 5 BIT0.








QCBOXOUTBYTE8-BIT binary value that will be output from Model 5 port.

QCBOXPHANTOMMicrophone power supply voltage (V) set for Model 5 IN1 and IN2 input.Ranges from 2 to 24.

8.4 EXTERNAL TRIGGER

It is possible to trigger the QC tests sequence with the following:1) A foot pedal switch connected to QCBox Pedal In connector.2) The connection of the loudspeaker under test sensed by QCBox Model 5.3) An external TTL signal wired to one of the QCBox Model 5 input.4) An external TTL signal wired to the PC parallel printer port.

The settings are within CLIO Options>QC.

This operation is controlled by the External Trigger button in the QC control paneland by the MANUAL keyword inside the QC script.

Figure shows a foot pedal switch and shows its connection to the PC to enable thecontrol of the QC test.

PCLPT Start

The QCBox Model4 and Model 5 have a dedicated input 'PEDAL IN' that can beused to connect the external foot pedal or trigger signal.

The following lines are needed inside a script file to enable a switch (or externallygenerated TTL signal) to start and continue a QC measurement.

[GLOBALS]......MANUAL=0

Please refer to later paragraphs and to the commands reference for more details onTTL input signal management.

8.5 TTL SIGNALS GENERATION

CLIO QC has powerful capabilities to generate and read TTL control signal to beable to interface with an external line automation.

To manage these TTL signals it is possible to use:1) The parallel port of the computer, if present.



2) The dedicated Digital I/O port of the QCBox Model 5 (USB controlled).

It is possible to define the status of the bits of the digital port involved; thefollowing is a list of the kind of signals possible:- signals output at startup (INITIALBITS, QCBOXINITIALBYTE)- signals conditioned by the result of a single measure ([IF LAST GOOD], [IF LASTBAD])- signals conditioned by the global result ([IF ALL GOOD], [IF ALL BAD])- unconditioned signals ([PERFORM])

Let's see an example of generation of external signals conditioned by the result ofthe measurement (LPT Parallel Port case):

[GLOBALS]......INITIALBITS=0[FFT]......[MLS]......[IF LAST BAD]BIT=3BITVALUE=1DELAY=200[IF LAST GOOD]BIT=3BITVALUE=0DELAY=200[IF ALL GOOD]BIT=1BITVALUE=1[PERFORM]BIT=0BITVALUE=1DELAY=200[PERFORM]8BITVALUE=0

This example defines a signal high on bit 3 if the MLS test performs bad, a signalhigh on bit 1 if all the tests are OK and an unconditioned pulse of 200 ms on bit 0that may be used to signal the end of the QC test sequence.

Referring to next figure we can see the time signal of the three bits in the twopossible cases A and B; in case A the MLS test performed bad and in case B good.



Let's see now translate the same script with QCBox Model 5 dedicated keywords:

[GLOBALS]......QCBOXINITIALBYTE=0[FFT]......[MLS]......[IF LAST BAD]QCBOXOUTBIT3=1DELAY=200[IF LAST GOOD]QCBOXOUTBIT3=0DELAY=200[IF ALL GOOD]QCBOXOUTBIT1=1[PERFORM]QCBOXOUTBIT0=1DELAY=200[PERFORM]QCBOXOUTBYTE=0



8.6 TIME DELAYS GENERATION

It is possible to define a time delay in any point of a script file with the followingdefinition:

[PERFORM]DELAY=200

In this example the QC sequence waits for 200 millisecond when encounteringthese keywords. In the previous paragraph you can also see the possibility ofmixing time delays with signals definitions in order to generate pulses.



8.7 PARALLEL PORT SIGNALS MANAGEMENT

The TTL signals generated with the active parallel printer port of the PC may beinteractively controlled by means of the QCBox&LPT menu recallable with Shift-F4.After opening this box press the Direct TTL Control button and you obtain thecontrol panel shown in figure. To get TTL signals operation please select a parallelport from the ones available.

The Direct TTL Controls dialog lets you set the status of the eight output bits usingthe appropriate check boxes while triggering it with the Set Bits button; a decimalrepresentation of the output binary word is also present. On the left side the statusof the input start bit is reported.

The pin-out of the standard parallel port is shown below; note the eight output bitsand the start trigger pulse in input.

BIT 1

BIT 2

BIT 3

BIT 4

BIT 5

BIT 6

BIT 7

START

BIT 0

11421531641751861972082192210231124122513



8.8 QCBOX MODEL 5 DIGITAL I/O SIGNALS MANAGEMENT

Using the QCBox Model 5 Digitali I/O port it is possible to generate/monitor TTLsignals to be used for interfacing along production lines; these key features arecontrolled over the USB connection of the Model 5 so there is no need for legacydevices like LPT ports.

The TTL signals generated with the QCBox Model 5 may be interactively controlledby means of the QCBox&LPT dialog recallable with Shift-F4. After opening this boxpress the Model 5 button and you obtain the control panel shown in figure.

The Model 5 Controls dialog lets you interactively set the status of the output bitswhile monitoring the input ones; simply click on green pin to control its status.

The pin-out of the Digital I/O port of the QCBox Model 5 is shown in figure; notethe six output bits, the four input bits and the +5V line.



8.9 RS-232 SERIAL PORT CONTROL

During QC execution it is possible to control serial devices, like label printers,connected via an RS-232 link to your PC. You can select and configure a COM portfor QC control within CLIO Options>QC.

The following script can be used to print a label at the end of a QC test if the resultof the test is good; the printing commands refer to a Zebra Z4M printer.

[GLOBALS]OPENSERIAL=1SERIALMONITOR=1..................[IF ALL GOOD]SERIALOUT=^XA^LH40,100,^F020,10^AD^FD@SERIALNUMBER^FS^XZ

Note the @SERIALNUMBER acronym that is used to output the current serialnumber. It is possible to activate, mainly for debugging purposes, a monitor windowthat echoes RS-232 activity; to do this use the SERIALMONITOR keyword.

The same text output in the above example could be saved in an ASCII file andloaded with the SERIALOUTFILE keyword:

...

...[IF ALL GOOD]SERIALOUTFILE=SERIAL.TXT



10 COMPLETE EXAMPLE: FAST, SINGLE-TEST LOUDSPEAKER QC

CLIO QC is able to program and execute a very fast, accurate and complete, single-test quality control of a loudspeaker using the new functionality of the sinusoidalmeasurement menu.

One of the key features of this approach is represented by the new FAST-TRACK™rub&buzz detection that is carried out along the sinusoidal sweep.

With just one sinusoidal sweep it is possible to measure:- Frequency response- Impedance response- Sensitivity- Polarity- Total harmonic distortion response- Single harmonic response (from 2nd to 10th)- Rub&Buzz- T&S parameters (Fs,Qt,Qe,Qm,Cms,Mms,Mmd,Vas,Bl,dBSPL,ZMin)

Choosing among the various settings of the sinusoidal test it is possible to tailor theQC test easily, controlling the trade-offs between speed and accuracy.

This example describes a test setup and the relative QC script that may beimplemented in an automatic production line capable of cycle times of 1 to 2seconds with a sweep time of around 1s.

10.1 HARDWARE REQUIRED

The following parts of the CLIO system are needed to achieve this kind of QC test:- CLIO FW-01- QCBox Model 5- Microphone (MIC-01,02 or 03)- Optional 19”rack QC panel

The CLIO system hardware presents itself as in this picture:

The basic connections required are listed here:1) On the electro-acoustic side we find the QCBox used as power amplifier,microphone directly connected to CLIO, current sensing to channel B input tomeasure impedance.2) On the digital side we find the connection with an external automation that givesa TTL start signal to the QC test and is informed by three output bits of its currentstatus.



ISense

DIGITAL I/O

In1 D.U.T.

QCBox Model 5

OUT A

CLIO

IN A OUT BIN B

Mic

Speaker

CLIOIn2 In3 In4To From

CLIO

1239BIT0 OUTBIT1 OUTBIT2 OUTBIT2 IN

To Automation

3) On the loudspeaker under test side a suitable acoustic test fixture should besetup to properly isolate from the outside environment. The example is not dealingwith this topic.

To properly control the QCBox Model 5 verify its settings within the QCBox&LPTControls dialog; default settings (i.e. 2A output current limiting) should be OK formany DUTs.

We will deal later with QCBox settings for connecting with the automation.

Once the hardware connections are firmly setup take a reference loudspeakerrepresentative of the production and put it in place ready to be measured. Wesuppose to deal with a wideband automotive 4” loudspeaker.

This quality control application relies on a stereo sinusoidal test that simultaneouslymeasures frequency response by means of a microphone connected to input A andimpedance sensing load current to input B. It is suggested to divide the initialapproach in two separate single channel measurements of the two quantities andfinally integrate them into a single stereo one.



10.2 MEASURING THE REFERENCE FREQUENCY RESPONSE

Open the sinusoidal menu. Let’s start with the acoustic frequency response and setup the required sweep opening the settings dialog. The main parameters affectingsweep are: frequency range chosen from 30Hz to 15kHz, resolution of 1/12 ofoctave supposed to be fine and speed that is set to “Fast” as best tradeoff forrub&buzz testing.

Before taking the first reference measurement you still need to set the properoutput level (here chosen 1V at speaker terminals) as indicated by DUTspecifications and accordingly set input sensitivity of CLIO input A; as the finalmeasurement will be stereo operate separately the two input channel controlsreleasing the Link Input Controls button in the hardware toolbar; initial input Asensitivity is -10dBV (channel B is left to 0dBV).

Now, within the sinusoidal menu, choose CHA input channel selection and dBSPL asY scale unit. Press go. The first measurement gives you the following result

one important parameter now clear is the sweep time that is shown in thesinusoidal menu status bar: with these settings we have 1.05 seconds sweep time.Consider it fine. Save the result to “response.sin” file.

The test should now be tuned up to take into account the acoustic environment andcompleted with missing settings. Open the sinusoidal settings dialog; proper delayshould be set to compensate for microphone distance to loudspeaker, this may beevaluated by the two common ways CLIO gives you i.e. taking a trial sinusoidalmeasurement with auto delay active or taking a parallel MLS&LogChirpmeasurement and inspecting the impulse response; in our case we found a 0.2msdelay to be compensated, due to a quasi near field measurement with amicrophone to DUT distance, in the acoustic fixture, of circa 7cm. Final settingsrequired are about distortion curves; we need to activate THD and Rub&Buzzcalculations clicking on “THD Enabled” and “R&B Enabled”, the Rise parameter is setto 0dB as we are going to accommodate all displayed curves inside one single100dB Y scale graph. Execute the measurement with final frequency responsesettings.



After the measurement is done we may inspect THD and Rub&Buzz pressing therelative buttons, in figure they are shown as overlays (green THD, red R&B). Repeatthe measurement until fully confident with the results obtained, eventually refinethe settings as needed.

We are now ready to define QC masks for frequency response, THD and Rub&Buzz.Open the QC menu, press the limits button to start defining a limit definition; werequire, and manually input, a relative mask to frequency response with thefollowing behavior:

[RELATIVE][UPPER LIMIT DATA]20 1080 10100 35000 36000 520000 5[LOWER LIMIT DATA]20 -1080 -10100 -35000 -36000 -520000 -5

The limits definition for THD and R&B is, for their nature, inherently absolute andonly requires an upper curve so we may begin defining them by direct, on-screen,drawing.



Pressing the THD button, inside sinusoidal menu, you obtain the THD curve; insideQC you press the “Draw Limits Controls” button and you are allow to draw a limitcurve directly on the sinusoidal graph; at the end the QC limit definition panel willbe filled with data about the drawn limit:

[THD UPPER LIMIT DATA]30.81 69.56174.96 69.78639.82 80.621603.03 78.1910869.90 64.2510869.90 64.25

Pressing the R&B button, inside sinusoidal menu, you obtain the R&B curve; insideQC you press the “Draw Limits Controls” button and you are allowed to draw a limitcurve directly on the sinusoidal graph; at the end the QC limit definition panel willbe filled with data about the drawn limit:

[RUB+BUZZ UPPER LIMIT DATA]30.11 48.32143.15 48.76445.16 60.491692.02 59.383924.35 30.18

It is time to save the limits file as “response.lim”.

As we are dealing with an unsmoothed frequency response that is presenting somehigh frequency “peaks and dips” we like to give a 1/6 of octave frequency jitteringto the calculated limits curve:

[RELATIVE]FREQJITTER=0.16

That gives our frequency mask a more comfortable behavior that is less prone togive false negatives in that troubled spectrum range.



The final parameter that we should take into account is the sensitivity of theloudspeaker; programming a chirp with same frequency extremes and analyzing itwith the multimeter we obtain a sensitivity of the reference of 106 dBSPL; thisvalue leads us to complete the limits file definition with sensitivity data.

[SENSITIVITY]UPPER=109LOWER=103

10.3 MEASURING THE REFERENCE IMPEDANCE RESPONSE

We put now our attention to the impedance response of our loudspeaker.

Going back to sinusoidal menu we choose CHB with the input channel selector andOhm as Y Scale unit; inside the sinusoidal settings dialog leave all previous settingsunchanged as they will accompany us to the final reference measurement; onlychange the impedance settings to “QCBox Select” to reflect QCBox operation.

As the output level has already been set for the acoustic test we only have to dealwith input sensitivity for channel B; a settings of -30dBV or -40dBV is usuallycorrect for ISense impedance tests. The measurement looks as follow.

Save the result to “impedance.sin” file.

We are now able to define also the limits file needed to check the impedanceresponse. Going back to the QC menu, inside limits control panel, we should clearinformation about frequency response masks and be ready for new input.

A response check mask may be defined as follow:

[RELATIVE]PERCENT=1[UPPER LIMIT DATA]20 2050 2060 3090 30100 20



200 201200 20[LOWER LIMIT DATA]20 -2050 -2060 -3090 -30100 -20200 -201200 -20

This 20% wide mask spans from below resonance to slightly higher the ZMin regionand opens up to 30% in resonance region.

The most important QC checks will be done on T&S parameters that take intoaccount all possible defects from the impedance point of view. In this definition wecheck Fs, Qms and ZMin to be within 10% from reference.

[TSPARAMETERS]PERCENT=1DIAMETER=10REDC=7KNOWNMMD=5FSUPPER=10FSLOWER=-10QMSUPPER=10QMSLOWER=-10ZMINUPPER=10ZMINLOWER=-10

It is time to save the limits file as “impedance.lim”.



10.4 INTEGRATING THE QC REFERENCE FILE

Starting from the actual situation, i.e. having just measured impedance relying onsettings that accumulated from the previous frequency response measurement, weare now ready to integrate all of our work to realize a single stereo sinusoidalmeasurement that will be the reference for our QC script.

Go to the sinusoidal menu, have the impedance measurement loaded in memory;select CHA&B with the input selector, change the Y Scale unit to dBSPL; CLIO isnow ready to take a two channels measurement with main unit set to dBSPL; as themeasured unit for channel B needs to be Ohm we must open the sinusoidal settingsdialog and select “Ohm Right Scale”: in this way channel B will measureimpedance using the right scale to identify it.

The final sinusoidal settings are:

Press Go; the graph obtained has frequency response measured from channel Aand refers to left scale while impedance response comes from channel B referringto right scale. Note that the two curves displayed are measured and controlled bydedicated checkboxes, no overlays are active.

This measurement is OK to be the reference for our QC test; once the frequencyscale, Y right and left scales are OK for the visualization under QC it can be savedas “reference.sin”. To properly set scales it is useful to directly input values at theirextremes; refer to 6.2 and 6.4 for details about this.



10.5 PROGRAMMING THE QC SCRIPT

We are ready to write the QC script; the files involved are the stereo sinusoidalmeasurement stored inside the “reference.sin” file, the limits file for channel A in“response.lim” and the limits file for channel B in “impedance.lim”.

[SIN]OUTQCBOX=1VINA=-10INB=-40REFERENCE=REFERENCE.SINLIMITSA=RESPONSE.LIMLIMITSB=IMPEDANCE.LIM

Two things are still missing: the polarity check and the visualization of the rub&buzzcurve.

1) To add polarity check over the frequency response we simply add POLARITY=1under [SIN]

2) To add rub&buzz display as a third curve together frequency response andimpedance we add [RUB+BUZZ DISPLAY] in the “response.lim” file.

The final script will thus be:

[SIN]OUTQCBOX=1VINA=-10INB=-30REFERENCE=REFERENCE.SINLIMITSA=RESPONSE.LIMLIMITSB=IMPEDANCE.LIMPOLARITY=1

This script may now be saved as “faststereosweep.qc”.

The final “response.lim” file will be:

[RELATIVE]FREQJITTER=0.16[SENSITIVITY]UPPER=109LOWER=103[UPPER LIMIT DATA]20 1080 10100 35000 36000 520000 5[LOWER LIMIT DATA]20 -1080 -10100 -35000 -36000 -520000 -5[THD UPPER LIMIT DATA]



30.81 69.56174.96 69.78639.82 80.621603.03 78.1910869.90 64.2510869.90 64.25[RUB+BUZZ DISPLAY][RUB+BUZZ UPPER LIMIT DATA]30.11 48.32143.15 48.76445.16 60.491692.02 59.383924.35 30.18

The final “imedance.lim” file will be:

[RELATIVE]PERCENT=1[UPPER LIMIT DATA]20 2050 2060 3090 30100 20200 201200 20[LOWER LIMIT DATA]20 -2550 -2560 -3090 -30100 -25200 -251200 -25[TSPARAMETERS]PERCENT=1DIAMETER=10REDC=7KNOWNMMD=5FSUPPER=10FSLOWER=-10QMSUPPER=10QMSLOWER=-10ZMINUPPER=10ZMINLOWER=-10



10.6 RUNNING THE QC TEST

Running the complete QC test we will obtain a comprehensive graph display as infigure.

10.7 ADDING THE INTERFACE TO AUTOMATION

To manage TTL signals that connect the system QCBox Model 5 digital I/O port tothe external automation we must include some programming inside CLIO and insidethe QC script.

As we have chosen input Bit 2 to trigger the QC test we must set this inside CLIOOptions>QC dialog.

The output bits operation should be defined directly inside the QC script and shouldreflect how CLIO and the automation interact.

We suppose the following meaning of the output TTL bits:BIT0 -> Signals the end of the sweep.BIT1 -> Signals if result is good.BIT2 -> Signals if result is bad.

The keywords that should be added to our script lead to the following situation

[PERFORM]



QCBOXOUTBYTE=0

[SIN]OUTQCBOX=1VINA=-10INB=-30REFERENCE=REFERENCE.SINLIMITSA=RESPONSE.LIMLIMITSB=IMPEDANCE.LIMPOLARITY=1

[PERFORM]QCBOXOUTBIT0=1

[IF ALL GOOD]QCBOXOUTBIT1=1

[IF ALL BAD]QCBOXOUTBIT2=1

Here you may see the initial keyword QCBOXOUTBYTE=0 that resets all threesignals to zero. Then, after the test is finished they are set to reflect the end ofsweep and the result of the test.

The only thing to be noted is that BIT0 (end of sweep) is output when thesinusoidal test is finished, i.e. right after all calculations and measurementmanagements are made. This means that it will be delayed with respect to theactual end of the sweep by the time the computer takes to make all the calculationsand actions related to a sinusoidal test; this time is usually small but not zero andmay range in some hundred of milliseconds depending on the platform chosen.

If very tight synchronization is needed and you want to avoid the calculation time,it is possible to require that the TTL signal is output right after the sweep iscompleted, without waiting for the sinusoidal test to end; to do this, place therelative keyword right under the [SIN] definitions thus changing the script to:

[PERFORM]QCBOXOUTBYTE=0

[SIN]OUTQCBOX=1VINA=-10INB=-30REFERENCE=REFERENCE.SINLIMITSA=RESPONSE.LIMLIMITSB=IMPEDANCE.LIMPOLARITY=1QCBOXOUTBIT0=1

[IF ALL GOOD]QCBOXOUTBIT1=1

[IF ALL BAD]QCBOXOUTBIT2=1

You can download these example files from Audiomatica website.



11 QC APPS

11.1 QC OF A MICROPHONE PREAMPLIFIER

Input

PreamplifierPRE-01

OUTPUT ACLIO

INPUT A

OUTPUT B

INPUT B

Output

This example is taken form our internal QC procedure for the PRE-01 MicrophonePreamplifier. Figure shows the connections required. The PRE-01 features threeweighting filters and two gain positions. This test is a representative case of thefollowing requirements:

1) The limits are ABSOLUTE as they are taken from the IEC tables for the specifiedtolerance. Since the perfect device has still to be built it is not possible to userelative limits from a real life measured reference.

2) The IEC specifies a response in term of a 0 dB at 1kHz. The absolute level at1kHz is however left to the test procedure. As we want to perform the test near thehighest level the device is able of accept, we need to use the PROCESS feature toshift the real measurement to the specs level.

3) Changes in switch position are required during test. We have therefore to usethe INTERACTIVE feature.

4) A level regulation is required to align the gain at 1kHz with and without a filter.This brings in the LOOP feature of the [MET] multimeter test.

5) It’s very difficult for the operator to set a switch accordingly to the next test tobe performed. The PERFORM and MESSAGE feature greatly simplifies this,avoiding errors.

The QC script, described here with comments, allows the check of the filtersresponse against Type 1 tolerance specification. It also checks for +/- 0.2 dB gaintolerance of the gain switch in both positions. As an additional feature it allows theuser, within the test, to adjust a variable gain trimmer that has to be adjusted toachieve optimum levels; this procedure, LOOP, also ends with a check of theadjusted level to be within +/- 0.2 dB. At every level check a distortion test, THDdefined in the LEV1.LIM file, is performed. As a general rule a QC procedure isdefined from one QC file (.qc extension) and several limits file (.lim extension)declared in the qc file. Process files (.mpro or .spro) are also involved here andthese are the only ones not specifically QC related. It is a good idea to dedicate adirectory for each QC test. The files involved here are:



PRE01.QC

LEV1.LIMA.LIMB.LIMC.LIMASHIFT.SPRO


[GLOBALS]COMPANY=AUDIOMATICA S.R.L. FLORENCETITLE=PRE01 TEST PROCEDUREINTERACTIVE=1SAVEONBAD=1

[PERFORM]MESSAGE=FILTER OFF DIP ON OFF OFF OFF

[MET]OUT=2.44IN=10REFERENCE=FILTER.METLIMITS=LEV1.LIM

[PERFORM]MESSAGE=FILTER ON DIP ON OFF OFF OFF

[MET]OUT=2.44IN=10REFERENCE=FILTER.METLIMITS=LEV1.LIMLOOP=1

[PERFORM]MESSAGE=FILTER ON DIP ON OFF OFF ON

[MET]OUT=-17.56IN=10REFERENCE=FILTER.METLIMITS=LEV1.LIM

[SIN]OUT=-10IN=10REFERENCE=A.SINLIMITS=A.LIMPROCESS=ASHIFT.SPRO

[PERFORM]MESSAGE=FILTER ON DIP OFF ON OFF ON

[SIN]OUT=-10IN=10REFERENCE=A.SIN



LIMITS=B.LIMPROCESS=ASHIFT.SPRO

[PERFORM]MESSAGE=FILTER ON DIP OFF OFF ON ON

[SIN]OUT=-10IN=10REFERENCE=A.SINLIMITS=C.LIMPROCESS=ASHIFT.SPRO

[PERFORM]MESSAGE=SET DEFAULT SETTINGS FILTER OFF DIP ON OFF OFF ON



11.2 THE AMPLIFIER&SWITCHBOX UNDER QC

ISense

From CLIOIn 1

D.U.T.

Ampli/SwitchBoxCLIOQC Model 4

OUTPUT ACLIO

INPUT A

OUTPUT B

INPUT B

To CLIO

In 2

5 Ohm 1%

This example details the quality control procedure that Audiomatica uses to test itsproduction of CLIOQC Amplifier & Switchbox.

A precision 5 Ohm 10W 1% resistor is needed and must be connected across DUTterminals. The procedure, executed in Interactive mode, guides the operator andrequests the manual connection of the unit; the cable coming from output B ofCLIO must be swapped during the test between input 1 and 2.

The test begins with two impedance measurements, the first executed in ISenseMode, the second executed in Internal Mode. Then a THD measurement with FFTand finally the frequency response of each input channel are performed.

Note the keywords used to alternatively mute CLIO's output.

[GLOBALS]COMPANY=AUDIOMATICA S.R.L. FLORENCETITLE=QCBOX TEST PROCEDUREINTERACTIVE=1

[PROMPT]MESSAGE=CONNECT:MESSAGE2=[OUTA->FROM CLIO][INA->TO CLIO][OUTB->CH1][INB->ISENSE]

[PROMPT]MESSAGE=PLACE 5 OHM 1% RESISTOR ACROSS D.U.T. TERMINALS

[SETIMPEDANCE][SETMUTEB]

[PERFORM]DELAY=500

[SIN]



OUT=0IN=-20REFERENCE=IMPEDANCE.SINILIMITS=IMPEDANCE.LIM

[SETINPUT1]

[PERFORM]DELAY=500

[SIN]OUT=10IN=-20REFERENCE=ISENSE.SINILIMITS=IMPEDANCE.LIM

[FFT]OUT=10.0IN=-10ACQUISITIONDELAY=200REFERENCE=FFT.FFTLIMITS=FFT.LIM

[RESETMUTEB][SETMUTEA]

[PERFORM]DELAY=500

[SIN]OUT=10IN=10REFERENCE=CH.SINLIMITS=CH.LIM

[PROMPT]MESSAGE=CONNECT:MESSAGE2=[OUTB -> CH2]

[SETINPUT2]

[PERFORM]DELAY=500

[SIN]REFERENCE=CH.SINLIMITS=CH.LIM

[RESETMUTEA]




11.3 A TEST ON A STEREO ELECTRONIC EQUIPMENT

The following self-explaining script implements the procedure required to test thefrequency response of a stereo equipment; it is simulated by a couple of PRE-01units each connected, as in the picture, to the two channels of CLIO. Both PRE-01have A-weighting filter active; the unit connected to channel B has +20dB gain.

Beyond the frequency response of the two channels, the script also measure theA/B difference response and output it to QC screen.

[SIN]OUT=0.0 dBVINA=-10INB=20REFERENCE=PRE01_A_B20.SINLIMITSA=AB_A.LIMLIMITSB=AB_B.LIM




11.4 A CYCLIC SCRIPT (USED TO MANAGE MY ROGERS LS3/5A TWO-WAY LOUDSPEAKER PRODUCTION)

This example describes a hardware and software setup to do quality control over aproduction of loudspeakers units; the responses are taken come from our samplesof Rogers LS3/5A speakers. The hardware setup is shown in figure

ISense

From CLIOMic No 1

Mic No 2

Speaker

Ampli/SwitchBox

far field

near field

CLIOQC

OUTPUT ACLIO

INPUT A

OUTPUT B

INPUT B

To CLIO

As you can see we employ a CLIOQC Amplifier & SwitchBox that connects twomeasuring microphones, one for near field response and the other for far fieldresponse. The internal switcher is used to configure impedance with current sensingor frequency response measurements and to select the correct microphone.

The quality control of such a production relies on what is called a referenceloudspeaker i.e. a unit which is kept aside the line and retested regularly to givereference data curves for the units under test. These data trace environmentalconditions.

To accomplish the recurrent operation of testing the reference loudspeaker CLIO QCimplements what is called the cyclic script i.e. a QC script that is launched by themain script on a timed basis and executed once. When the cyclic script is launchedthe operator is prompted and the reference unit must be placed on the line.

The three keywords used to define this operation are CYCLIC, REPETITION andCYCLICFIRST under [GLOBALS]. CYCLIC defines the name of the cyclic script; thisfile must reside in the same directory of the calling one. REPETITION defines afterhow many units it is run; we put 4 in the example only to allow you to test it, thisnumber is chosen after evaluating the particular condition of the production line.CYCLICFIRST, which in the example is commented away, tells the software toexecute the cyclic script before the first run of the main script; this is useful to setknown conditions at the beginning of a QC session.

[GLOBALS]



CYCLIC=ROGERSCYCL.QC;CYCLICFIRST=1REPETITION=4OUTUNITS=VOUTQCBOX=2.83IN=-20

Please note the use of the OUTUNITS keyword which accounts for output levelsexpressed in Volts RMS. With OUTQCBOX=2.83 we chose to set 2.83 Volts atRogers terminals.

The rest of the main script for producing my LS3/5As deals with the three actualmeasurements for testing nearfield, farfield and impedance data; the first two aredone with MLS, the third with Sinusoidal. Before each measurement definition arethe relative commands that set the correct function of the Amplifier & SwitchBox;note that the impedance is done in 'ISense' mode.

[SETINPUT1]

[MLS]REFERENCE=NEARFIELD.MLSLIMITS=NEARFIELD.LIM

[SETINPUT2]

[MLS]REFERENCE=FARFIELD.MLSLIMITS=FARFIELD.LIM

[SETISENSE]

[SIN]OUTQCBOX=1IN=-30REFERENCE=IMPEDANCE.SINILIMITS=IMPEDANCE.LIM

The main QC script ends here. It is a fairly simple one, which can be customized forany production of loudspeakers. Let's now see the cyclic script. The basic idea is toexecute the same measurements as in the main script and save them with thenames of the reference files for the main script itself. AUTOSAVE=1 prepares forsaving all the measurements done; SAVEFOLDER= is a particular syntax to set thescript directory as the current one.

[GLOBALS]AUTOSAVE=1SAVEFOLDER=OUTUNITS=VOUTQCBOX=2.83IN=-20

The rest of the cyclic script resembles the main script with the difference that aftereach measurement, we define the name of the file to be saved and force it to beequal to the name of the reference file; in this way the reference file itself isupdated. SAVEPROMPT=1 instructs the QC processor to prompt for user acceptanceof the save operation; this is useful for validating the procedure and avoidingerrors.



[SETINPUT1]

[MLS]REFERENCE=NEARFIELD.MLSLIMITS=NEARFIELD.LIMSAVENAME=NEARFIELDSAVEPROMPT=1

[SETINPUT2]

[MLS]REFERENCE=FARFIELD.MLSLIMITS=FARFIELD.LIMSAVENAME=FARFIELDSAVEPROMPT=1

[SETISENSE]

[SIN]OUTQCBOX=1IN=-30REFERENCE=IMPEDANCE.SINILIMITS=IMPEDANCE.LIMSAVENAME=IMPEDANCESAVEPROMPT=1




11.5 QC OF A TELEPHONE WITH LOUDNESS RATING CHECK

We describe here the quality control test of the production of a telephone unit.

Two script files take part to this example.

The cyclic script “moutheq.qc” is needed to measure and save the output pressureresponse of the reference speaker or mouth at reference position.

[GLOBALS]AUTOSAVE=1SAVEFOLDER=[PERFORM]MESSAGE=PLACE REFERENCE MICROPHONE IN PLACE[SIN]OUT=-28 dBuIN=-10REFERENCE=MOUTH.SINLIMITS=NONESAVENAME=MOUTH

The main script “phone.qc” tests the frequency response of the phone under testequalizing the drive pressure at -4.7 dBPa; also defined inside the “phone.lim”limits file a Send Loudness Rating QC check.

[GLOBALS]CYCLIC=MOUTHEQ.QCCYCLICFIRST=1REPETITION=100[PERFORM]MESSAGE=PLACE TELEPHONE IN PLACE[SIN]OUT=-4.7IN=-10EQREFERENCE=MOUTH.SINREFERENCE=REFPHONE.SINLIMITS=PHONE.LIM




11.6 ON RUB & BUZZ DETECTION (1)

This example describes an effective technique to detect rub&buzz in a productionline of loudspeakers. The technique is based on logarithmic chirp stimulus withsynchronous FFT detection.

CLIO is able to generate (see 7.7) logarithmic chirps of proper length and properstart and stop frequencies.

Given your production of speakers you should program the log chirp following theseguidelines:Frequency Range. The frequency extremes depend on the kind of speaker; thestart frequency must be below the resonant frequency (Fs) to achieve excursionwhile the stop frequency should be high enough to stimulate all possible defectsand anomalous mechanical contacts. We suggest start to lie between 20Hz/100Hzwhile stop between 500Hz/1500Hz. Stop should be a compromise between bestdefect detection and anomalous resonances excitation.

Amplitude. Perhaps this is the most critical parameter to set. Its choice must takeinto consideration T&S parameters of the device and tend to exploit the maximumexcursion possible (XMax). On the other side a too high stimulus amplitude willtend to give false positives to R&B. The graph below shows excursion normalizedversus Qt and Fs; it tells us that, in free air (as it is usually the case of productionlines), maximum excursion is reached well below Fs (around 0.1*Fs).

This leads us also to consider the technique described after (19.9.9) to apply DCand relax other parameters while augmenting R&B detection.Duration. It is directly related to the chirp length; at 48 kHz sampling you getthe following: a 16k chirp lasts around 0.35s, a 32k chirp lasts around 0.7s, a64k chirp lasts around 1.4s and so on. The choice should be consistent with your production test needs provided a longertest should be preferable as some kind of R&B phenomena appear with time asdevice thermal constants are reached. For the same reason if R&B is one amongother QC tests, it should be done at the end.



Once the stimulus has been defined you must define a proper FFT QC test; be sureto use the same size of the stimulus, i.e. FFT Size = Chirp Size. Anotherimportant FFT parameter to set is smoothing which will present an easier to detectanalysis; we suggest 1/48 or 1/24th of octave smoothing.

The analysis leads to the following situation:

You can see the response of a good and a rubbing device which will lead you tocorrect mask definition; it is also shown how this measurement detects theharmonic signature of the device; the plateau marked with 2nd directly refers tosecond harmonic response.

This QC test is as simple as the following definition:

[FFT]COMMENT=RUB&BUZZQCBOXDCOUT=2.83IN=0REFERENCE=RUB.FFTLIMITS=RUB.LIM

We set 2.83V at the QCbox output (given a former OUTUNITS=V definition) andinput at 0dBV. Extreme care must be put in order to optimize inputsensitivity as this measurement is very sensitive to noise.

Limits mask should be placed in the decaying part of the acquisition and extendendto cover the highest frequencies; only upper limit is necessary in this case.



11.7 ON RUB & BUZZ DETECTION (2)

This example describes a simple method to enhance rub&buzz detection. Thismethod is based on the possibility of applying a DC voltage superimposed to thegenerated stimulus. This technique applies to any test possible with CLIO andaugments its sensitivity.

As it is evident also from the figures in 19.9.7 the maximum excursion is obtainedat DC and this is an effective way to bring the speaker to its limits. As it is evidentfrom the following figure when a DC is applied the corresponding AC signalamplitude must be lowered to obtain similar excursion.

Applying a DC to the same QC test as described before in 19.9.8 it is possible toobtain the following measurement where it is evident the much better detection ofthe defect which is possible.

As described in 4.6.1 it is possible to manually set the DC voltage at the output of aQCBox Model 5 using the relative control panel.



Under a QC script it is possible to apply DC with the following synthax:

[PERFORM]QCBOXDCOUT=1.2[SIN]REFERENCE=RESPONSE.SINLIMITS=RESPONSE.LIM[PERFORM]QCBOXDCOUT=-1.2[SIN]REFERENCE=RESPONSE.SINLIMITS=RESPONSE.LIM[PERFORM]QCBOXDCOUT=0

In this example it has been applied a 1.2V DC voltage to a sinusoidal test; thesame could have been applied to a FFT with log chirp or any other test; to benoted that the same test must be executed twice as we don’t know a prioriwhich direction stimulates the defect to arise.

In this case also lower harmonics could be checked as, when DC is present, theybecome sensitive to R&B too.



11.8 A C++ CLIENT APPLICATION TO CONNECT TO TCP/IP SERVER

A fully commented sample C++ client console application that is able to connect toCLIO, request measurements and receive results follow:

/* clio client c - code for example client program that uses TCP */

#include <windows.h>#include <winsock.h>#include <stdio.h>#include <string.h>

#define PROTOPORT 1234 /* default protocol port number */

extern int errno;char localhost[] = "localhost"; /* default host name *//*------------------------------------------------------------------------ * Program: clioclient * * Purpose: allocate a socket, connect to the Clio Server, interact with * the QC environmet * * Syntax: client [ host [port] ] * * host - name of a computer on which server is executing * port - protocol port number server is using * * Note: Both arguments are optional. If no host name is specified, * the client uses "localhost"; if no protocol port is * specified, the client uses the default given by PROTOPORT. * *------------------------------------------------------------------------ */int string_length(char str[]);

main(argc, argv)int argc;char *argv[];{ struct hostent *ptrh; /* pointer to a host table entry */ struct protoent *ptrp; /* pointer to a protocol table entry */ struct sockaddr_in sad; /* structure to hold an IP address */ int sd; /* socket descriptor */ int port; /* protocol port number */ char *host; /* pointer to host name */ int n; /* number of characters read */ char ibuf[100]; /* buffer for data from the server */ char obuf[100]; /* buffer for data to the server */

WSADATA wsaData; WSAStartup(0x0101, &wsaData);

memset((char *)&sad,0,sizeof(sad)); /* clear sockaddr structure */ sad.sin_family = AF_INET; /* set family to Internet */

/* Check command-line argument for protocol port and extract */ /* port number if one is specified. Otherwise, use the default */ /* port value given by constant PROTOPORT */

if (argc > 2) { /* if protocol port specified */ port = atoi(argv[2]); /* convert to binary */ } else { port = PROTOPORT; /* use default port number */ } if (port > 0) /* test for legal value */ sad.sin_port = htons((u_short)port); else { /* print error message and exit */ fprintf(stderr,"bad port number %s\n",argv[2]); exit(1); }

/* Check host argument and assign host name. */



if (argc > 1) { host = argv[1]; /* if host argument specified */ } else { host = localhost; }

/* Convert host name to equivalent IP address and copy to sad. */

ptrh = gethostbyname(host); if ( ((char *)ptrh) == NULL ) { fprintf(stderr,"invalid host: %s\n", host); exit(1); } memcpy(&sad.sin_addr, ptrh->h_addr, ptrh->h_length);

/* Map TCP transport protocol name to protocol number. */

if ( ((int)(ptrp = getprotobyname("tcp"))) == 0) { fprintf(stderr, "cannot map \"tcp\" to protocol number"); exit(1); }

/* Create a socket. */

sd = socket(PF_INET, SOCK_STREAM, ptrp->p_proto); if (sd < 0) { fprintf(stderr, "socket creation failed\n"); exit(1); }

/* Connect the socket to the specified server. */

if (connect(sd, (struct sockaddr *)&sad, sizeof(sad)) < 0) { fprintf(stderr,"connect failed\n"); exit(1); }

/* Wait a little */

n=0;while (n < 1000000) {n=n++;}

/* Get greeting message */

n = recv(sd, ibuf, sizeof(ibuf), 0);write(1,ibuf,n);

/* Repeatedly read write data from socket or stdin and write to user's screen. */

while (strcmp(obuf,"exit\n")) { fgets(obuf,127,stdin); n = send(sd, obuf, string_length(obuf), 0); n = 0; while (n < 1000000) {n = n++; }

n = recv(sd, ibuf, sizeof(ibuf), 0); write(1,ibuf,n);

}

/* Close the socket. */

closesocket(sd);

/* Terminate the client program gracefully. */

exit(0);}



int string_length(char str[]){ int i; for(i = 0; i < 80; i++) {

if(str[i] == '\0') { return(i); }

}}





Date post:	04-May-2018
Category:	Documents
Upload:	vandung
View:	218 times
Download:	1 times

CLIO QC EXPLAINED WITH APPS - Audiomatica · CLIO QC EXPLAINED WITH APPS ... The QC software is a...

Documents